Direct electrocatalytic epoxidation of olefins: advances in membrane electrode assemblies and beyond

Abstract

Electrocatalytic epoxidation of olefins represents a promising and sustainable pathway for producing high-value epoxides, such as propylene oxide. This review comprehensively examines recent advancements in catalyst design and membrane electrode assembly (MEA) reactor engineering, while also addressing persistent challenges including catalyst cost, stability, and mass transfer limitations. Although MEA technologies have achieved remarkable progress, exemplified by an over 25% reduction in energy consumption, their industrial deployment remains constrained by issues such as Nafion membrane degradation and inefficient transport of long-chain olefins. Future research endeavors should prioritize the development of cost-effective, durable catalytic systems and their seamless integration with renewable energy sources to facilitate the large-scale implementation of green electrochemical epoxidation processes.

---

Yuzheng Li

Yuzheng Li is currently pursuing a doctoral degree in chemistry at the College of Science, Tianjin University, under the guidance of Professor Li, focusing on the research of organic electrochemical transistors.

Hui Li

Hui Li is currently pursuing a Master's degree in Chemical Engineering and Technology at China University of Petroleum (East China), under the guidance of Professor Li, focusing on the research of electrocatalytic direct epoxidation of propylene.

Yinghua Zhang

Yinghua Zhang is currently pursuing a Master's degree in Chemical Engineering at China University of Petroleum (East China), under the guidance of Professor Li, focusing on the research of electrocatalytic direct epoxidation of ethylene.

Ruiji Wang

Ruiji Wang is currently pursuing a Master's degree in Chemical Engineering at China University of Petroleum (East China), under the guidance of Professor Li, focusing on the research of electrocatalytic acetylene coupling.

Zhongtao Li

Zhongtao Li is a distinguished professor at the School of Science, Tianjin University, and a professor at the College of Chemical Engineering, China University of Petroleum (East China). He has been granted the Shandong Province “Outstanding Young Scientist” fund and recognized as a “Taishan Scholar”. His current research focuses on the structural design and development of nanocomposite materials in the fields of new energy and catalysis.

1. Introduction

Olefins are essential chemical feedstocks, predominantly derived from petroleum cracking. Their epoxides serve as versatile intermediates in the production of fine chemicals, food additives, pharmaceuticals, and agrochemicals. The global market for propylene oxide (PO), for example, exceeded 15 million tons in 2024, with an annual growth rate of 3.5%, highlighting its industrial significance.^1–4

The evolution of olefin epoxidation reflects a broader shift toward greener chemical technologies. Initial electrocatalytic efforts date back to the 1960s, when researchers began replacing traditional oxidants with electrical energy. Early studies revealed that anodic oxidation could generate reactive intermediates, such as hydroxyl and peroxo species to drive epoxidation.⁵ However, these systems were constrained by low efficiency and poor selectivity due to suboptimal electrode materials (e.g., Pt, Ru). By the 1990s, advances in carbon-based composites (e.g., graphene, carbon felt) combined with metal oxides had significantly enhanced electrode conductivity and stability.⁶ Moreover, precise electrolyte regulation (e.g., pH control via buffer solutions) and dynamic potential management effectively suppressed undesirable side reactions, such as over-oxidation or C–O cleavage thereby leading to notable improvements in selectivity and yield.^7,8 The emergence of green chemistry further underscored the environmental benefits of electrocatalysis, which eliminates the need for highly toxic oxidants like peracids and chromates.⁹

Since 2010, significant progress has been made in developing novel catalytic materials, from single-atom catalysts to metal–organic frameworks (MOFs), which offer precise active sites for epoxidation.^10–12 However, a critical bottleneck remains in the reactor configuration. Traditional aqueous H-cells are severely hindered by the low solubility of gaseous olefins and large ohmic resistance (>10 Ω), limiting current densities to levels insufficient for industrial application (<10 mA cm⁻²). To overcome these physical constraints, membrane electrode assembly (MEA) reactors have emerged as a transformative solution.^13–15 By integrating gas diffusion electrodes (GDEs) in a zero-gap configuration, MEAs establish a robust triple-phase interface, minimizing mass transfer resistance and enabling ampere-scale electrolysis suitable for industrial scale-up.^16,17 Central to the performance of MEA reactors is the control of catalytic pathways. Direct electrocatalytic epoxidation (DEP), the focus of this review, relies on the anodic activation of water to generate reactive oxygen species. The selectivity of this process is fundamentally governed by the competitive adsorption of three key intermediates: atomic oxygen (O*), hydroxyl radicals (OH*), and peroxo species (OO*). While the O* pathway (common on noble metals) favors direct oxygen addition to the C=C bond, it risks competitive deep oxidation. In contrast, OH* or OO* pathways (typical of transition metal oxides) offer milder oxidation routes but are highly sensitive to the local proton/hydroxide concentration.¹⁸

Despite these advances, a clear knowledge gap exists in bridging microscopic active sites with macroscopic reactor engineering. Current literature often treats catalyst design and reactor optimization as separate entities. In contrast to previous reviews, this work introduces a ‘cross-scale interfacial coupling’ framework.^19,20 We argue that the dominance of a specific pathway (O*/OH*/OO*) is dictated not only by the catalyst's intrinsic electronic structure but also by the dynamic microenvironment (e.g., local pH, hydrophobicity) within the MEA. This review systematically examines these coupling effects, identifies the ‘mismatch’ between material stability and reactor conditions, and provides a roadmap for overcoming the barriers to commercialization.

With the rapid expansion of renewable electricity generation from solar and wind sources, electrons have emerged as a clean reagent for organic synthesis. As conceptually illustrated in Fig. 1, DEP couples this renewable energy with ubiquitous water molecules to convert olefins into high-value epoxides. This process operates under ambient conditions, avoids hazardous reductants or oxidants, and allows precise control over reaction progress through regulation of voltage and current, offering higher yields and selectivity. Thus, electrochemical epoxidation is recognized as a green and efficient route to epoxide production.^21,22

Fig. 1 The influence of optimizing interface mass transfer and accelerating interface reaction on direct epoxidation.

Electrocatalytic olefin epoxidation can be classified into indirect and direct pathways. As illustrated in Fig. 2a and b, indirect electrooxidation employs redox mediators (e.g., Cl⁻) to generate active species (e.g., HOCl) that epoxidize olefins. While sometimes selective, this approach involves multi-step reactions, reducing energy efficiency and introducing stability issues due to mediator decomposition and corrosion.^23–26 In contrast, DEP involves the reaction of olefins at the anode with oxygen species derived from water. The selectivity of DEP is fundamentally governed by the competitive adsorption of active intermediates (O*, OH*, and OO*) on the catalyst surface.^27,28 This method eliminates explosion risks associated with mixing oxidants (e.g., H₂O₂, BrO⁻) with organic substrates and simplifies product separation. Therefore, DEP of olefins is of great scientific and practical interest.^29–32

Fig. 2 (a) Schematic of chloride-mediated electrochemical propylene epoxidation setup. (b) Schematic of membrane electrode device for water-mediated electrochemical propylene epoxidation.

Catalyst selection and optimization are critical, as these reactions require mild conditions and high selectivity. While recent reviews have extensively summarized the electronic structure of electrocatalysts and general organic oxidation mechanisms, they predominantly focus on half-cell reactions in liquid electrolytes.^2,3 Consequently, a critical gap remains in bridging microscopic active sites with macroscopic reactor engineering. Distinct from previous works, this review introduces a “cross-scale interfacial coupling” framework within the context of MEA reactors. We argue that the performance of direct epoxidation is not solely determined by the catalyst's intrinsic activity but is governed by the dynamic microenvironment within the MEA.³³ We shift the perspective from atomic-site design to system-level integration, specifically addressing the compatibility between catalyst layers and ion-exchange membranes, water management at the triple-phase interface, and strategies to overcome mass transfer limits in zero-gap configurations. This review uniquely elucidates how MEA components—ranging from the ionomer distribution in the catalyst layer to the pore gradient in gas diffusion electrodes—modulate the local pH, water activity, and reactant concentration. By correlating these engineering parameters with the competitive adsorption of reactive oxygen species (O*, OH*, OO*), we provide a transformative perspective on overcoming the persistent trade-offs between selectivity, energy efficiency, and stability under industrially relevant conditions. The following sections review advances in reactor design, catalytic systems, and interfacial engineering for electrocatalytic epoxidation, with a specific emphasis on MEAs and future directions.³⁴

2. Electrocatalytic epoxidation reactors

2.1. Traditional electrocatalytic epoxidation reactors

The performance of electrocatalytic olefin epoxidation depends not only on catalyst properties but also critically on reactor design, which governs current density, mass transfer, and energy efficiency. Commonly used reactors include H-cells, flow reactors, and MEA reactors.

The H-cell represents a classic electrochemical setup and remains ubiquitous in fundamental laboratory research. As shown in Fig. 3a, it consists of separate anode and cathode compartments separated by an ion-exchange membrane to prevent electrolyte crossover. In a typical epoxidation setup, the working and reference electrodes are positioned in the anode chamber, where gaseous olefin is sparged into the electrolyte, while the counter electrode is located in the cathode compartment.^35,36 The inherent simplicity and compartmentalization of H-cells ensure operational stability, while the inclusion of a reference electrode facilitates precise potential control. Consequently, H-cells are indispensable for preliminary catalyst screening and mechanistic investigations across various organic electrosynthesis reactions, including liquid olefin epoxidation, alcohol oxidation, and carboxylic acid reduction.³⁷

Fig. 3 Evolution of reactor configurations. (a) Conventional H-cell, strictly limited by low gaseous olefin solubility and large ohmic resistance. (b) Flow cell incorporating a gas diffusion electrode. (c) Membrane Electrode Assembly reactor.

However, H-cells exhibit significant limitations. For gaseous olefins like propylene, which are typically introduced via bubbling, mass transfer constraints lead to concentration polarization and low current densities. The rough surface morphology of conventional electrodes further reduces Faraday efficiency.^38,39 Additionally, as batch reactors, H-cells require frequent electrolyte and catalyst replacement, and their small reaction scale makes them unsuitable for industrial applications.

Liquid flow reactors, depicted in Fig. 3b, typically consist of three independent flow channels for gas, catholyte, and anolyte.⁴⁰ A key feature is the use of a GDE to separate the anolyte channel from the olefin gas stream. The catalyst layer is deposited on the electrolyte-facing side of the GDE, while olefin is delivered from the gas side, facilitating reaction at the catalyst surface. This design enhances mass transfer efficiency and maximizes the catalyst-reactant contact area, significantly improving performance.⁴¹ The GDE establishes a gas–liquid-solid triple-phase interface, allowing gaseous olefins to react directly without prior dissolution, thereby mitigating mass transfer limitations.^42,43

For example, researchers evaluated Ag₃PO₄ catalysts with different morphologies for propylene electrooxidation in a GDE-equipped flow cell.⁴⁴ Cubic Ag₃PO₄ achieved nearly 80% PO selectivity and a partial current density of 0.49 mA cm⁻² at 2.4 V (vs. RHE), corresponding to a production rate of 5.3 g(PO)m⁻² h⁻¹—a record performance among reported electrocatalysts. In another study, a bromide-mediated ethylene epoxidation system maintained a Faraday efficiency of 80–90% across various current densities, with an ethylene oxide production rate exceeding 1 kg m⁻² h⁻¹ at 156 mA cm⁻².³⁷ The system also demonstrated stable operation for 4.5 hours without degradation. Similarly, Wang et al. explored halogen-mediated epoxidation of propylene on a RuO₂/Ti anode, achieving over 80% Faraday efficiency, and up to 90% in bromide-mediated systems at low potentials (<1.5 V).⁴⁵

Despite these improvements, flow reactors still face challenges: catalyst current density remains relatively low, aqueous electrolytes lead to product dilution and complicated separation, and the presence of water exacerbates side reactions like the oxygen evolution reaction (OER).^46–49

2.2. MEA reactors: advantages and challenges

MEA reactors represent a major advancement in electrocatalytic epoxidation technology (Fig. 3c). Early systems, which used phosphoric acid-soaked glass wool separators and Pt black catalysts, achieved only modest propylene oxide production rates (37 µmol h⁻¹ cm⁻²) and Faradaic efficiencies (7.4%) due to gas solubility and mass transfer limitations.⁵⁰

Recent innovations in GDEs and MEAs have addressed these issues. As shown in Fig. 4, GDEs feature a gradient porous structure-comprising a current collector, diffusion layer, and catalyst layer-that establishes a micrometer-scale triple-phase interface.⁵¹ Gradient porous electrodes utilize a structural design where pore size gradually decreases from the gas diffusion layer to the catalyst layer, aiming to balance gas transport and electrolyte wetting.⁵² This allows gaseous olefins to access catalytic sites directly, bypassing dissolution and greatly enhancing mass transfer.⁵³ The catalyst layer requires careful balance between hydrophilicity and hydrophobicity: excessive hydrophobic additives (e.g., PTFE) can block ion and electron transport channels, while over-wetting may flood pores and block access to active sites. Optimized ionomer coverage (3–5 nm) combined with a gradient hydrophobic substrate ensures proton conduction without obstructing the diffusion of larger olefin molecules.⁵⁴

Fig. 4 Hierarchical gradient structure design and interface optimization of gas diffusion electrodes in MEA reactors. Reproduced with permission from ref. 44, Copyright 2022 by Springer Nature.

MEA technology further reduces interelectrode distance to the hundred-micrometer scale using ultrathin proton exchange membranes (PEMs), reducing ohmic losses by approximately two orders of magnitude.⁵⁵ An effective MEA requires synergistic integration of the catalyst layer, PEM, and gas diffusion layer (GDL). The catalyst layer must maximize active site exposure and mass transfer kinetics, the PEM should possess high sulfonic acid group density for proton conductivity and an ultrathin reinforced structure to resist swelling and radical attack.⁵⁶ Additionally, the GDL should employ bio-inspired fractal pore designs and gradient fluorination to balance gas transport and water management, preventing cathode flooding and anode drying.⁵⁷

Nevertheless, MEA systems still face challenges related to long-term stability and adaptability: local hot spots at high current densities cause noble metal sintering, chemical degradation leads to the loss of sulfonic acid groups in Nafion membranes. Furthermore, the conversion efficiency of long-chain olefins is often constrained by significant steric hindrance and slower diffusion kinetics within the micropores of the catalyst layer compared to smaller propylene molecules.^58–63

Addressing these issues requires innovations at the material, structural, and system levels. Material-level strategies include developing core–shell catalysts to modulate oxygen intermediate adsorption via shell oxygen vacancies, thus inhibiting deep oxidation and enhancing sintering resistance. Ultrathin graphene-modified composite membranes with 3D cross-linked networks can mitigate thermal expansion mismatch and suppress delamination. Structurally, reinforced composite membranes should combine high sulfonic acid density with radical scavenging capabilities, while molecular imprinting on catalyst surfaces could enable directional olefin adsorption. At the system level, intelligent thermal management and dynamic interface layers are needed to maintain a stable reaction environment. These dynamic layers must accommodate catalyst surface remodeling occurring at the electrode–electrolyte interface or the real-time evolution of ion concentration within the electric double layer at reaction potentials.^64,65

Despite these structural advances, a critical limitation persists in current MEA designs: the ‘hydrophobicity-conductivity paradox’. Most strategies rely on heavy PTFE loading to prevent flooding, which inadvertently disrupts the ionomer network and drastically increases ohmic resistance. This creates a bottleneck where mass transfer is improved at the expense of energy efficiency. Furthermore, unlike liquid-phase flow cells, current MEA configurations lack effective heat dissipation mechanisms, leading to local hotspots that accelerate membrane degradation. Future iterations must therefore prioritize the decoupled control of gas and ion transport channels.⁶⁶ In the long term, MEA technology is expected to evolve into integrated precision reactors that enhance mass transfer, selectivity, and lifetime, providing an efficient and stable platform for green electrocatalytic olefin epoxidation. These developments will not only enable the efficient synthesis of chemicals like propylene oxide but also advance electrochemical synthesis toward industrial-scale application, supporting the chemical industry's transition under global carbon neutrality goals. The optimization of the MEA is a multi-parameter engineering challenge.⁶⁷ As detailed in Table 1, the interplay between the gas diffusion substrate, membrane type, and ionomer distribution is critical for balancing mass transfer with ionic conductivity. Successful designs often employ gradient structures to manage the complex three-phase interface.⁶⁸

Table 1 Key design parameters and optimization strategies in MEA reactors

Reactor component	Optimization strategy	Key parameters/materials	Effect/benefit	Ref.
GDE substrate	Gradient porosity	Carbon paper/Ti mesh	Enhances gas transport & regulates wetting	51 and 57
Catalyst layer	Hydrophobicity control	PTFE/ionomer ratio	Prevents flooding of active sites	69
Membrane	Thickness reduction	Ultrathin nafion/PEM	Reduces ohmic resistance (RO mega) by 2 orders	55 and 70
Interface	Cross-linking	Graphene oxide-amino acid	Improves thermal stability & ion selectivity	71
Ionomer	Loading optimization	3–5 nm coverage	Balances proton conduction vs. gas blockage	54 and 72

---

Despite the structural advances in MEA configurations, a critical limitation persists in current designs: the “hydrophobicity-conductivity paradox”. Most optimization strategies rely on increasing the polytetrafluoroethylene PTFE loading in the gas diffusion electrode to prevent flooding. However, this hydrophobic treatment inadvertently disrupts the ionomer network, creating insulating regions that drastically increase the local ohmic resistance. Consequently, mass transfer is improved often at the expense of energy efficiency and voltage stability. Furthermore, unlike liquid-phase flow cells that naturally dissipate heat, current zero-gap MEA configurations lack effective heat dissipation mechanisms, leading to local hotspots that accelerate membrane degradation and catalyst detachment. Future reactor engineering must therefore prioritize decoupled control strategies, such as developing hydrophilic–hydrophobic dual-gradient active layers, to simultaneously ensure rapid gas transport and efficient ion conduction without mutual interference.

3. MEA reactor optimization: catalyst selection

During electrochemical oxidation, active species such as ˙O, ˙OH, and ˙OO can be generated to oxidize olefins to epoxides.^70–73 As shown in Fig. 5a and b, in the direct epoxidation reaction of propylene, the O-related pathway relies on O* dissociated and adsorbed on the catalyst surface to directly attack the propylene double bond. The reaction is as follows, where the generation of O* is the rate determining step, and the enrichment of propylene gas and diffusion of O* are primary mass transfer limitations. Structural optimization involves establishing micropores and hydrophobic layers, and the capillary action of micropores increases the local concentration of propylene, Hydrophobic layer prevents water molecules from blocking O* active sites.

H₂O → O* + 2H⁺ + 2e⁻ (1)

C₃H₆ + O* → C₃H₆O (2)

The OH coordination pathway activates propylene through hydrogen bonding with OH*, which reduces the energy barrier of epoxidation and significantly improves selectivity. However, precise control of moisture content is required. The reaction is as follows, where the formation of OH* is the rate determining step, and water molecule transport and H⁺ supply are the primary mass transfer limitations. Structural optimization involves employing mesoporous architectures and ionomer coatings. Specifically, mesopores facilitate electrolyte wetting and ensure sufficient H₂O transport. A thin layer of ionomer provides proton conduction channels, and the water content must be carefully adjusted to balance the coverage of OH* and the reaction energy barrier.

H₂O → OH* + H⁺ + e⁻ (3)

C₃H₆ + OH* → [C₃H₆⋯OH]* (4)

[C₃H₆⋯OH]* → C₃H₆O + H⁺ + e⁻ (5)

The OO coordination pathway utilizes OO* to synergistically insert oxygen into double bonds, combining the advantages of high selectivity and low energy barrier. The reaction is as follows, where the generation of OO* is the rate-determining step involving multi-electron transfer and the formation of high-energy peroxide species. However, long-chain olefins encounter high diffusion resistance in micropores, leading to a decrease in OO* pathway selectivity. Consequently, the contact efficiency of OO represents the primary mass transfer bottleneck. Structural optimization is achieved by establishing macropores and gradient hydrophobic substrates. Specifically, macropores serve as the main gas channels to accelerate gas transport, while the hydrophobic substrate prevents long-chain olefins from blocking OO* active centers.

2H₂O → OO* + 4H⁺ + 4e⁻ (6)

C₃H₆ + OO* → C₃H₆O + O* (7)

Fig. 5 (a) Scheme for the reaction pathway between O and OH. (b) The plan for OO reaction pathway.

3.1. Noble metal catalysts

As shown in Fig. 6a, precious metals such as Pt, Pd, and Ag exhibit O binding energies and activities near the summit of the volcano plot for epoxidation reactions. The unique advantage of precious metal catalysts lies in their ability to optimize intermediate adsorption energy through atomic-level electronic structure regulation, thereby enabling precise control of the olefin epoxidation reaction pathway. As shown in Fig. 6b, the Dai's team investigated the DEP mechanism of atomic oxygen and molecular oxygen on group IB metal surface (direct epoxidation of propylene), and compared key competitive reactions, including the formation of oxametallacycle (OMMP) intermediates and the Allylic Hydrogen Stripping (AHS) reaction.⁷⁴ A novel peroxametallacycle intermediate (OOMMP) was identified, which can be further converted into the OMMP intermediate via O–O bond cleavage. When considering both OOMMP formation and O–O bond cleavage processes, the effective energy barrier for the molecular oxygen (O₂*) species during the entire OMMP formation process is only 0.26 eV, which is significantly lower than the AHS energy barrier. This indicates that the O₂* pathway achieves superior selectivity for OMMP intermediates. Conversely, in the atomic oxygen mechanism, the AHS reaction proceeds with lower activation energy compared to OMMP formation, suggesting that overall PO selectivity is likely to be poor when atomic O* serves as the oxidant.⁷⁵

Fig. 6 (a) Volcano activity models of electrolytic propylene epoxidation plotted with different weak-binding material surfaces at 1.60 VRHE. Reproduced with permission from ref. 76, Copyright 2022 by the American Chemical Society. (b) Reaction profiles for OMMP formation (black curve) and the AHS reaction (red curve) in the molecular oxygen mechanism on the Ag (111) surface. The effective barrier (Eeffa) for OMMP formation is depicted by the dashed blue curve, and the results obtained on Cu and Au surfaces are listed in the bottom-left corner. Reproduced with permission from ref. 74, Copyright 2017 by the Royal Society of Chemistry. (c) Free energy diagrams for propylene epoxidation on Ag–V–O/GDL and Ag–O/GDL (insert: energy barriers for the epoxidation between *O and *C₃H₆ on Ag–V–O/GDL and Ag–O/GDL. *OMC: *CH₃–CH(CH₂)–O). Reproduced with permission from ref. 72, Copyright 2025 by Springer Nature. (d) Free energy diagrams of propylene epoxidation with molecular oxygen through the one-step mechanisms. Reproduced with permission from ref. 72, Copyright 2025 by Springer Nature. (e) Proposed reaction scheme for propene oxidation in water on the clean Pd fcc (111) surface and 6*CO poisoned Pd fcc (111). Reproduced with permission from ref. 29, Copyright 2022 by the Royal Society of Chemistry.

As shown in Fig. 6c, Li et al. developed Ag/V bimetallic catalysts.⁷² To elucidate the role of Ag–V–O/GDL catalysts (containing Ag–O species) in promoting the reaction, the Gibbs free energy (ΔG) of the proposed reaction pathway was calculated using density functional theory (DFT). The formation of *O is identified as the rate-determining step (RDS). Compared with Ag–O/GDL, the RDS (*O formation) on Ag–V–O/GDL in the electrochemical propylene epoxidation process exhibits a lower ΔG (0.805 eV vs. 0.954 eV). Furthermore, the energy barrier for *O generation on Ag–V–O/GDL is 1.569 eV, which is significantly lower than that on Ag–O/GDL (2.770 eV), indicating that *O is more easily generated on the Ag–V–O surface. The incorporation of V helps form Ag–O active centers, thereby reducing the formation energy of *O radicals. Meanwhile, V doping shifts the d-band center of Ag sites downwards. Consequently, the coupling of *O with adsorbed propylene to form the key intermediate (*OC₃H₆) is accelerated, leading to a significant increase in PO yield. As shown in Fig. 6d, Li et al. studied a pathway identified on the silver surface through OMMP.⁷⁶ On Au, an element with weak oxygen binding ability, O=O activation and conversion from OMMP to PO have relatively high potential barriers (1.07 eV and 0.96 eV, respectively). Therefore, there may be other DEP mechanisms that are more suitable for its weak binding properties. Through extensive search of other transition states using the climbing-image nudged elastic band method, a one-step mechanism was discovered (propylene directly picks up atomic oxygen and forms PO), which significantly reduces the TS barrier (0.07 eV), indicating that there may be a more favorable mechanism on weak oxygen binding surfaces such as Au (100), where the direct PO formation barrier is much lower than the OMMP mechanism. As shown in Fig. 6e, Chorkendorff's team⁶⁵ used in situ spectroscopy to reveal a potential reaction pathway bifurcation mechanism, at potentials >1.1 V (vs. RHE), the palladium surface oxidizes to form an active PdO₂ stage, activating H₂O to form a highly active OOH intermediate (adsorption energy ΔE = −1.12 eV), which preferentially attacks the C=C bond of propylene to form epichlorohydrin, At a potential of <1.1 V, the weak adsorption property of Pd^δ+–O bond, leads to a decrease in OOH stability (ΔE = −0.85 eV), causing the reaction pathway to shift towards the cleavage of the intramolecular C–H bond of propylene, forming acrolein. Manthiram's team designed a Pd–Pt core–shell structure, in which Pt²⁺ is embedded in the PdO lattice, significantly suppressing the H₂O dissociation side reaction and increasing the PO selectivity to 66%.

3.2. Non-noble metal catalysts

The innovation of nonprecious metal catalysts lies in opening up new pathways for olefin epoxidation through atomic size distribution and synergistic effects of multiple metals. As shown in Fig. 7a, Huang's team calculated the ΔG_*OOH for each configuration based on the calculated hydrogen electrode (CHE) model.⁷⁷ The optimal adsorption free energy was determined to be 4.22 eV. The Co₂N₄O₂ model exhibited the most favorable electrocatalytic activity, with its *OOH binding energy (4.19 eV) approaching the peak of a volcano. In contrast, the CoN₄ model showed a lower ΔG_*OOH of 4.08 eV, indicating stronger adsorption of *OOH. The overpotential required for the formation of *OOH on Co₂N₄O₂ is 0.03 V, which is significantly lower than the overpotential on CoN₄ (0.14 V). This low overpotential is superior to other cobalt-based structures (such as CoN₃O, CoN₂O₂, and CoNO₃) and other O-bridge bimetallic centers (M₂N₄O₂, where M = Fe, Co, Ni, Cu, Pd). As shown in Fig. 7b, the study of Mn₃O₄ revealed the core mechanism of direct oxygen supply by water molecules through the development of nanoparticles by Manthiram's team on manganese-based catalysts.⁷⁸ Isotope labeling experiments confirmed that lattice oxygen, derived from water activation, is the active species. Mn³⁺ active centers drive H₂O cleavage to generate the key oxygen intermediate O* (adsorption energy ΔE = −0.78 eV) via a dynamic redox cycle (Mn³⁺ ↔ Mn⁴⁺), achieving 42% Faraday efficiency in cyclooctene epoxidation. However, Mn³⁺ is prone to surface reconstruction causing activity decay. To address this, the team constructed an Ir–MnO_x heterointerface.⁴⁸ As shown in Fig. 7c, Ir is anchored via Ir–O–Mn coordination, inducing Mn⁴⁺ → Ir³⁺ electron transfer. This stabilizes Mn³⁺ while optimizing the O* generation path (ΔE reduced to −0.95 eV), increasing propylene oxide selectivity to 46 ± 4%. Breakthroughs in molecular-level Mn catalysts further enhance electronic modulation. Zhang's team designed a manganese porphyrin complex (TMP) MnCl.⁷⁹ The axial Cl⁻ ligand modulates the d-orbitals of the Mn center, generating the highly active intermediate [(TMP)Mn^vO]⁺. This intermediate forms a π-antibonding adsorption with the styrene C=C bond via an empty d_z² orbital (bond angle compressed to 105°), significantly reducing the epoxidation energy barrier (activation energy from 0.85 eV to 0.51 eV), achieving 89% epoxidation selectivity at 0.5 mol% loading.

Fig. 7 (a) Calculated activity volcano plot for a two-electron ORR. Reproduced with permission from ref. 77, Copyright 2024 by the American Chemical Society. (b) Proposed Mechanism for Electrochemical Epoxidation by Manganese Oxide Nanoparticles. Reproduced with permission from ref. 78, Copyright 2019 by the American Chemical Society. (c) Proposed mechanism of the electrochemical epoxidation. Reproduced with permission from ref. 78, Copyright 2019 by the Royal Society of Chemistry. (d) Electrocatalytic epoxidation of cyclooctene on Co₄P₂W₁₈ by directly using water as oxygen source. Reproduced with permission from ref. 80, Copyright 2024 by the Elsevier.

Cobalt-based catalysts expand the reaction pathway through multi-metal interfacial synergy. As shown in Fig. 7d, He et al. synthesized Co₄P₂W₁₈ clusters.⁸⁰ The Co–O–W heterointerface creates an electron-rich domain (Co center charge density increased by 0.25 e⁻). In situ Raman spectroscopy confirmed that terminal O sites trigger O–H bond homolysis at 1.6 V potential to generate O* (adsorption energy ΔE = −1.05 eV). Adjacent Co sites fix cyclooctene in an η²-C adsorption configuration (C=C bond length extended from 1.34 Å to 1.41 Å), driving the C–O bond formation energy barrier down to 0.42 eV (50% lower than single-metal CoO_x systems). The charge buffering effect of the phosphotungstate skeleton effectively stabilizes the Co²⁺/Co³⁺ oxidation state, maintaining >90% catalyst activity after 7 cycles. These results systematically reveal the regulatory logic of “coordination microenvironment intermediate adsorption → epoxidation path”, laying the foundation for low-cost, highly selective olefin epoxidation systems.

Despite breakthroughs in reaction mechanisms, the industrialization of non-noble metal catalysts is hindered by two core challenges, insufficient stability of active sites and imbalance between mass transfer and electron conduction. On one hand, molecular catalysts suffer from poor solution stability, TMP MnCl activity decreased by 35% after 12 hours of continuous operation due to Mn²⁺ leaching caused by β-H elimination of the porphyrin ligand. Solutions include confining active centers within rigid porous frameworks or designing chelating ligands to inhibit ligand dissociation via strong covalent bonds. On the other hand, multi-metal systems face high component leaching risk, Co₄P₂W₁₈ clusters exhibited Co leaching rates of 0.7 ppm h⁻¹ in acidic electrolyte (pH = 2). XPS analysis indicated surface Co–O–W bond breakage leading to W site exposure (W 4f binding energy shift of 0.8 eV). To provide a holistic view of the current materials landscape, Table 2 summarizes the key performance metrics and mechanistic features of representative catalysts. A comparison reveals that while noble metals generally offer higher intrinsic selectivity via the O* pathway, non-noble systems are closing the gap through concerted proton–electron transfer mechanisms, albeit often requiring higher overpotentials.⁵⁴

Table 2 Performance comparison of representative electrocatalysts for olefin epoxidation

Catalyst type	Electrolyte	Active species	Potential/current	Selectivity (%)	Faradaic Eff. (%)	Ref.
Noble metals
Cubic Ag3PO4	0.5 M KHCO3	O*/lattice O	2.4 V vs. RHE	—	∼80%	44
Ag–V–O/GDL	Alkaline	O*	—	High (PO)	High	72
Non-noble metals
Co2N4O2	Acidic	*OOH	—	0.03 V	High	77
Mn3O4	Neutral	Lattice O (from H2O)	—	—	42%	78
(TMP)MnCl	—	[(TMP)MnO]+	—	—	89%	79
Co4P2W18	pH = 2	O*	1.6 V	—	>90% (Activity)	80

---

Mass transfer and reactor adaptability defects also limit catalytic potential. Low cross-interface diffusion efficiency of water molecules in the Mn₃O₄ system restricts O intermediate generation. Future efforts should focus on, optimizing intermediate adsorption energy by altering metal center charge distribution, designing novel polyoxometalate-based carriers inspired by the phosphotungstate charge buffering mechanism to suppress active component leaching via strong covalent bonding, developing membrane-free electrolyzers adapted to non-noble metal catalytic characteristics, utilizing high-curvity microchannels to enhance mass transfer efficiency. Particularly noteworthy is the cross-fertilization of Mn-based and Co-based systems, Combining Mn's unique water activation capability with Co's coordination flexibility to construct Mn–Co dual-nuclear active sites holds promise for overcoming the intrinsic limitations of single-metal center electronic structures, providing a cross-scale solution for transitioning non-noble metal catalysts from lab validation to industrial application.⁸¹

A critical comparison of noble and non-noble systems reveals a distinct trade-off between selectivity and stability across these catalytic systems. Noble metals (e.g., Ag, Au) generally favor the direct O* pathway, offering superior intrinsic selectivity for propylene oxide, yet they are prone to surface reconstruction and sintering under the acidic conditions typical of high-performance MEAs. In contrast, non-noble metal oxides (e.g., Mn, Co) offer a cost-effective alternative by utilizing the lattice oxygen mechanism, but they often suffer from higher overpotentials and severe metal leaching.⁸² Consequently, the field is currently at an impasse: while noble metals define the upper limit of performance, the industrial viability of MEA electrolyzers likely hinges on stabilizing non-noble metals via lattice confinement or heterojunction engineering.

It is crucial to note that the intrinsic selectivity of these catalysts, determined in ideal H-cells, often degrades in practical devices. The MEA microenvironment drastically alters the adsorption energy of intermediates. For instance, a hydrophilic GDE surface may promote excessive water coverage, shifting the pathway from the desired O*-mediated epoxidation to the parasitic OER, regardless of the catalyst choice. Thus, catalyst design must be coupled with the interfacial engineering of the reactor.

A critical comparison of the literature reveals a distinct trade-off between intrinsic selectivity and operational stability across different catalytic systems. Noble metal catalysts, particularly Ag-based materials, generally favor the O* pathway, offering superior intrinsic selectivity for epoxide formation. However, they are prone to surface reconstruction and sintering under the acidic conditions typical of high-current MEA electrolyzers. In contrast, non-noble metal oxides (e.g., Mn, Co) offer a cost-effective alternative by utilizing lattice oxygen mechanisms or hydroxyl-mediated pathways, but they often suffer from higher overpotentials and severe metal leaching during long-term operation. Consequently, the field is currently facing a material bottleneck: while noble metals define the upper limit of catalytic performance, the industrial viability of MEA technology likely hinges on stabilizing non-noble metals via lattice confinement or heterojunction engineering to achieve a balance between cost, activity, and durability.

4. Industrial relevance and techno-economic considerations

The current industrial landscape for PO production is dominated by the chlorohydrin process and the hydrogen peroxide-to-propylene oxide (HPPO) process. The chlorohydrin route, while mature, suffers from generating substantial saline wastewater and toxic byproducts. The HPPO process is greener but relies heavily on the cost and transport safety of high-concentration H₂O₂. In contrast, DEP offers a disruptive alternative by generating the oxidizing species in situ from water. This eliminates the need for storing hazardous oxidants and theoretically simplifies the process flow. However, to displace existing technologies, DEP must match the unparalleled space-time yields of thermal catalysis.^83–85

From a techno-economic perspective, the economic viability of DEP is primarily dictated by the operational expenditure, where electricity consumption accounts for over 60% of the total cost. Preliminary techno-economic analyses (TEA) suggest that for electrochemical PO production to compete with the HPPO process, the system must achieve a Faradaic efficiency exceeding 90% at cell voltages below 2.5 V. The integration of MEA reactors plays a pivotal role here; by minimizing ohmic resistance, MEAs significantly reduce the energy input per kilogram of PO. Furthermore, coupling the process with low-cost renewable electricity is essential to bridge the cost gap with fossil-fuel-driven routes.

Despite the promise, transitioning from laboratory H-cells to industrial electrolyzers faces significant hurdles. (1) Current density gap: while academic studies often celebrate FEs at 10–50 mA cm⁻², industrial relevance requires stable operation at >300 mA cm⁻² to minimize capital expenditure on stack size. (2) Stability mismatch: industrial catalysts must endure thousands of hours of operation, whereas most DEP reports demonstrate stability for less than 100 hours. The degradation of anion exchange membranes and catalyst shedding under vigorous gas evolution remain unsolved issues. (3) Downstream separation: although MEAs reduce electrolyte volume, separating dilute PO (boiling point 34 °C) from the gas stream and unreacted propylene requires energy-intensive distillation, which must be optimized to ensure the overall system efficiency.

5. Conclusions

In the field of electrochemical olefin epoxidation technology, synergistic optimization of MEA reactors and catalytic systems has achieved breakthrough progress, yet cross-scale innovation is needed to overcome industrialization bottlenecks. MEAs integrate ultrathin PEMs and graded GDEs, compressing the electrode gap to the hundred-micrometer scale and reducing ohmic impedance by two orders of magnitude. Combined with solid-state electrolytes to suppress OER, they achieve low voltages and >25% energy reduction. However, long-term stability is limited by catalyst sintering at high current density, Nafion membrane sulfonic acid group loss, and selectivity decay for long-chain olefins.

Future perspectives and industrial outlook to bridge the gap between laboratory research and industrial application, future efforts must shift from trial-and-error material screening to rational system engineering. We propose three priority directions: (1) Standardization of stability protocols: current stability tests (often <100 hours) are insufficient for industrial benchmarking. The community should adopt accelerated stress tests that mimic fluctuation conditions of renewable energy inputs to accurately predict catalyst lifetimes over 10 000 hours. (2) Techno-economic analysis guided design: research must move beyond reporting only Faradaic efficiency. Metrics such as single-pass conversion rate, product concentration, and downstream separation costs must be integrated into the early stage of catalyst design. Developing catalysts that can operate efficiently at high reactant utilization rates (>50%) is crucial to reducing the energy penalty of product separation. (3) Precision microenvironment control: instead of focusing solely on the catalyst surface, attention should turn to the ionomer-catalyst interface. Developing specialized ionomers with high gas permeability and oxidative resistance will be key to unlocking the full potential of MEA electrolyzers. In summary, direct electrocatalytic epoxidation is poised for a breakthrough. By addressing the “cross-scale” challenges—from atomic active sites to macroscopic reactor heat management—this technology can evolve from a promising academic concept into a cornerstone of sustainable green chemistry.

Conflicts of interest

There are no conflicts to declare.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.

Acknowledgements

This research was financially supported by the National Natural Science Foundation of China (52403355), the China Postdoctoral Science Foundation (BX20240188 and 2024M751616).

Notes and references

1. X. Liu, X. Shen, H. Li, P. Li, L. Luo, H. Fan, X. Feng, W. Chen, X. Ai, H. Yang and Y. Cao, Adv. Energy Mater., 2021, 11, 2003905.
2. P. Zhang, T. Wang and J. Gong, Chin. Chem. Soc., 2023, 5, 1028–1042.
3. J. Teržan, M. Huš, B. Likozar and P. Djinović, Am. Chem. Soc., 2020, 10, 13415–13436.
4. P. Kube, J. Dong, N. S. Bastardo, H. Ruland, R. Schlögl, J. T. Margraf, K. Reuter and A. Trunschke, Nat. Commun., 2022, 13, 7504.
5. Q. Wang, L. Lyu, X. Hu, W. Fan, C. Shang, Q. Huang, Z. Li, Z. Zhou and Y. M. Kang, Angew. Chem., Int. Ed., 2025, 57, 6018–6041.
6. L. Wan, M. Pang, J. Le, Z. Xu, H. Zhou, Q. Xu and B. Wang, Nat. Commun., 2022, 13, 7959.
7. P. Sebastián-Pascual and M. Escudero-Escribano, ACS Energy Lett., 2020, 5, 130–135.
8. Z. He, J. Hwang, Z. Gong, M. Zhou, N. Zhang, X. Kang, J. W. Han and Y. Chen, Nat. Commun., 2022, 13, 3777.
9. Z. Deng, Y. Guo, Z. Sun, J. Lin, H. Zhai and W. Chen, Nano Res., 2024, 17, 9326–9344.
10. Z. Pei, H. Zhang, Y. Guo, D. Luan, X. Gu and X. W. Lou, Adv. Mater., 2023, 36, 2306047.
11. Y. Xiong, W. Sun, P. Xin, W. Chen, X. Zheng, W. Yan, L. Zheng, J. Dong, J. Zhang, D. Wang and Y. Li, Adv. Mater., 2020, 32, 200896.
12. Q. Wang, L. Lyu, X. Hu, W. Fan, C. Shang, Q. Huang, Z. Li, Z. Zhou and Y. M. Kang, Angew. Chem., Int. Ed., 2025, 64, e202422920.
13. A. Tombesi and C. Pettinari, Inorganics, 2021, 9, 81.
14. Q. Han, B. Qi, W. Ren, C. He, J. Niu and C. Duan, Nat. Commun., 2015, 6, 10007.
15. A. M. Viana, F. Leonardes, M. C. Corvo, S. S. Balula and L. Cunha-Silva, Materials, 2023, 16, 3133.
16. R. Haaring, P. W. Kang, Z. Guo, J. W. Lee and H. Lee, Acc. Chem. Res., 2023, 56, 2595–2605.
17. S. D. Shingte, O. Altenburg, P. J. T. Verheijen, H. J. M. Kramer and H. B. Eral, Cryst. Growth Des., 2022, 22, 4072–4085.
18. X. Guo, R. T. Gao, S. Ren, N. T. Nguyen, H. Chen, L. Wu and L. Wang, Nat. Commun., 2025, 16, 6220.
19. G. Cai, Y. Yin, D. Xia, A. A. Chen, J. Holoubek, J. Scharf, Y. Yang, K. H. Koh, M. Li, D. M. Davies, M. Mayer, T. H. Han, Y. S. Meng, T. A. Pascal and Z. Chen, Nat. Commun., 2021, 12, 3395.
20. S. Chai, S. He, Z. Li, Y. Zhu and H. Li, Chem. Eng. J., 2025, 15, 163174.
21. A. Winiwarter, L. Silvioli, S. B. Scott, K. Enemark-Rasmussen, M. Sariç, D. B. Trimarco, P. C. K. Vesborg, P. G. Moses, I. E. L. Stephens, B. Seger, J. Rossmeisl and I. Chorkendorff, Energy Environ. Sci., 2019, 12, 1055–1067.
22. C. Chen, Y. Li and P. Yang, Joule, 2021, 4, 737–742.
23. D. Song, B. Qiao, X. Wang, L. Zhao, X. Li, P. Zhang, Y. Yao, H. Chen, L. Zhu and H. Sun, Environ. Sci. Technol., 2023, 57, 9416–9425.
24. J. K. Liu, M. Kang, K. Huang, H. G. Xu, Y. X. Wu, X. Y. Zhang, Y. Zhu, H. Fan, S. R. Fang, Y. Zhou, C. Lian, P. F. Liu and H. G. Yang, Nat. Commun., 2025, 16, 3458.
25. Y. Zhang, J. Cui and Y. Pei, Chem. Eng. J., 2023, 464, 142652.
26. J. Wang, X. Dong, G. Feng, X. Lu, G. Wu, G. Li, S. Li, J. Mao, A. Chen, Y. Song, J. Zeng, W. Wei and W. Chen, Angew. Chem., Int. Ed., 2024, 63, e202411173.
27. J. S. Jirkovský, M. Busch, E. Ahlberg, I. Panas and P. Krtil, J. Am. Chem. Soc., 2011, 133, 5882–5892.
28. F. Dorchies, A. Serva, D. Crevel, J. De Freitas, N. Kostopoulos, M. Robert, O. Sel, M. Salanne and A. Grimaud, J. Am. Chem. Soc., 2022, 144, 22734–22746.
29. M. Ko, Y. Kim, J. Woo, B. Lee, R. Mehrotra, P. Sharma, J. Kim, S. W. Hwang, H. Y. Jeong, H. Lim, S. H. Joo, J. W. Jang and J. H. Kwak, Nat. Catal., 2022, 5, 37–44.
30. K. Dong, Y. Wang, L. Zhang, X. Fan, Z. Li, D. Zhao, L. Yue, S. Sun, Y. Luo, Q. Liu, A. A. Alshehri, Q. Li, D. Ma and X. Sun, Green Chem., 2022, 24, 8264–8269.
31. P. Ran, A. Qiu, T. Liu, F. Wang, B. Tian, B. Xiang, J. Li, Y. Lv and M. Ding, Nat. Commun., 2024, 15, 8877.
32. M. Chi, J. Ke, Y. Liu, M. Wei, H. Li, J. Zhao, Y. Zhou, Z. Gu, Z. Geng and J. Zeng, Nat. Commun., 2024, 15, 3646.
33. B. Zhou, R. Gao, J. J. Zou and H. Yang, Small, 2022, 27, 2202336.
34. X. Lin, Z. Zhou, Q. Y. Li, D. Xu, S. Y. Xia, B. L. Leng, G. Y. Zhai, S. N. Zhang, L. H. Sun, G. Zhao, J. S. Chen and X. H. Li, Angew. Chem., Int. Ed., 2022, 35, e202207108.
35. L. Cui, M. Sun and Z. Zhang, Green Chem., 2023, 25, 6315–6326.
36. W. Ju, F. Jiang, H. Ma, Z. Pan, Y. B. Zhao, F. Pagani, D. Rentsch, J. Wang and C. Battaglia, Adv. Energy Mater., 2019, 9, 1901514.
37. J. Schalck, J. Hereijgers, W. Guffens and T. Breugelmans, Chem. Eng. J., 2022, 446, 136750.
38. X. Huang, M. Song, J. Zhang, T. Shen, G. Luo and D. Wang, Nano-Micro Lett., 2023, 15, 86.
39. J. Yang, H. Yan, Q. A. Zhang, Y. Song, Y. Li and A. Tang, Small, 2023, 20, 2307354.
40. H. Rabiee, L. Ge, X. Zhang, S. Hu, M. Li and Z. Yuan, Energy Environ. Sci., 2021, 14, 1959–2008.
41. E. H. Majlan, D. Rohendi, W. R. W. Daud, T. Husaini and M. A. Haque, Renewable Sustainable Energy Rev., 2018, 89, 117–134.
42. L. Han, W. Zhou and C. Xiang, ACS Energy Lett., 2018, 3, 855–860.
43. R. Shi, J. Guo, X. Zhang, G. I. N. Waterhouse, Z. Han, Y. Zhao, L. Shang, C. Zhou, L. Jiang and T. Zhang, Nat. Commun., 2020, 11, 3028.
44. J. Ke, J. Zhao, M. Chi, M. Wang, X. Kong, Q. Chang, W. Zhou, C. Long, J. Zeng and Z. Geng, Nat. Commun., 2022, 13, 932.
45. J. Wang, G. Wu, G. Feng, G. Li, Y. Wei, S. Li, J. Mao, X. Liu, A. Chen, Y. Song, X. Dong, W. Wei and W. Chen, ACS Omega, 2023, 8, 46569–46576.
46. A. A. Ryan, S. D. Dempsey, M. Smyth, K. Fahey, T. S. Moody, S. Wharry, P. Dingwall, D. W. Rooney, J. M. Thompson, P. C. Knipe and M. J. Muldoon, Org. Process Res. Dev., 2023, 27, 262–268.
47. M. H. Guan, H. N. Xu, J. Liu, X. Y. Zhou, T. Wu and A. H. Lu, J. Am. Chem. Soc., 2025, 147, 16301–16308.
48. M. Chung, K. Jin, J. S. Zeng, T. N. Ton and K. Manthiram, J. Am. Chem. Soc., 2022, 144, 17416–17422.
49. M. Saito, Y. Kojima, S. Iguchi and I. Yamanaka, ACS Appl. Mater. Interfaces, 2024, 16, 67545–67552.
50. S. Iguchi, M. Kataoka, R. Hoshino and I. Yamanaka, Catal. Sci. Technol., 2022, 12, 469–473.
51. L. Li, W. Ye, Q. Liu, R. Liu, X. Lu, T. Yao, L. Wang, B. Gu, L. Sun and W. Yang, Proc. Natl. Acad. Sci. U. S. A., 2025, 122, e2418144122.
52. C. F. Toncón-Leal, J. Villarroel-Rocha, M. T. P. Silva, T. P. Braga and K. Sapag, Adsorption, 2021, 27, 1109–1122.
53. R. Zhao, L. Li, Q. Wu, W. Luo, Q. Zhang and C. Cui, Nat. Commun., 2024, 15, 8367.
54. L. Wang, R. Du, Z. Zhao, M. Na, X. Li, X. Zhao, X. Wang, Y. A. Wu, S. Jana, Y. Zou, H. Chen and X. Zou, Angew. Chem., Int. Ed., 2025, 64, e202501744.
55. S. Zhang, Z. Wang, R. Zhang, Y. He and K. Cen, Int. J. Hydrogen Energy, 2023, 48, 35463–35476.
56. K. Sato, T. Kajita and A. Noro, ACS Appl. Polym. Mater., 2023, 5, 3480–3488.
57. J. Zhang, W. Li, T. Ahmed Shifa, J. Sun, C. Jia, Y. Zhao and Y. Cui, J. Power Sources, 2019, 439, 227066.
58. G. Zhao, J. Lin, M. Lu, L. Li, P. Xu, X. Liu and L. Chen, Nat. Commun., 2024, 15, 8463.
59. R. Wan, T. Yuan, L. Wang, B. Li, M. Liu and B. Zhao, Nat. Catal., 2024, 7, 1288–1304.
60. H. He, S. Song, L. Zhai, Z. Li, S. Wang, P. Zuo, Y. Zhu and H. Li, Angew. Chem., Int. Ed., 2024, 63, e202409006.
61. W. Li, Y. Li, Z. Liu, H. Zhang, F. Jiang, B. Liu, Y. Xu, A. Zheng and X. Liu, Ind. Eng. Chem. Res., 2022, 61, 7760–7776.
62. D. Xiong, Z. Lai, J. Chen, H. Yuan and H. Wang, J. Am. Chem. Soc., 2025, 147, 17303–17314.
63. Z. Zhou, K. Zhang, P. He, H. Chang, M. Zhang, T. Yan, X. Zhang, Y. Li and Z. Cao, Angew. Chem., Int. Ed., 2024, 64, e202419900.
64. Z. Zhu, X. Yang, J. Liu, M. Zhu and X. Xu, Carbon Energy, 2023, 10, e327.
65. A. Donnadio, R. Narducci, M. Casciola, F. Marmottini, R. D’Amato, M. Jazestani, H. Chiniforoshan and F. Costantino, ACS Appl. Mater. Interfaces, 2017, 9, 42239–42246.
66. Y. Sun, J. Wang, S. Sun, H. Zhong, K. Wei, H. Liu, Q. Wei, Y. Li and F. Yang, Green Chem. Eng., 2024, 4, 431.
67. B. Li, X. Y. Wang, J. Li, L. Wang and F. L. Yu, Mol. Catal., 2025, 584, 115263.
68. G. P. Bein, M. A. Stewart, E. A. Assaf, S. J. Tereniak, R. N. Sampaio, A. J. M. Miller and J. L. Dempsey, ACS Energy Lett., 2024, 9, 1777–1785.
69. Z. Wang, Y. Li and Q. Li, Higher Education Press Limited Company, 2025, vol. 19, pp. 599–618.
70. Q. Wang, K. Sang, C. Liu, Z. Zhang, W. Chen, T. Ji, L. Li, C. Lian, G. Qian, J. Zhang, X. Zhou, W. Yuan and X. Duan, Nat. Commun., 2024, 15, 3249.
71. J. Liu, S. Wang, R. Yang, L. Li, S. Liang and L. Chen, J. Membr. Sci., 2022, 659, 120745.
72. Y. Lin, H. Li, X. Miao, Y. Sun, H. Ren, X. Yu, W. Cui, M. Wu and Z. Li, Nat. Commun., 2025, 16, 3113.
73. G. Y. Park, Y. S. Noh, H. Y. Jeong, S. J. Yoon, K. H. Oh, S. So, J. Kim, J. Choi and D. M. Yu, Eur. Polym. J., 2025, 234, 114024.
74. Y. Dai, Z. Chen, Y. Guo, G. Lu, Y. Zhao, H. Wang and P. Hu, Phys. Chem. Chem. Phys., 2017, 19, 25129–25139.
75. S. Möhle, M. Zirbes, E. Rodrigo, T. Gieshoff, A. Wiebe and S. R. Waldvogel, Angew. Chem., Int. Ed., 2018, 130, 6124–6149.
76. H. Li, C. S. Abraham, M. Anand, A. Cao and J. K. Nørskov, J. Phys. Chem. Lett., 2022, 13, 2057–2063.
77. H. Huang, M. Sun, S. Li, S. Zhang, Y. Lee, Z. Li, J. Fang, C. Chen, Y. X. Zhang, Y. Wu, Y. Che, S. Qian, W. Zhu, C. Tang, Z. Zhuang, L. Zhang and Z. Niu, J. Am. Chem. Soc., 2024, 146, 9434–9443.
78. K. Jin, J. H. Maalouf, N. Lazouski, N. Corbin, D. Yang and K. Manthiram, J. Am. Chem. Soc., 2019, 141, 6413–6418.
79. H. Wu, Y. Xu, P. Guo, Y. Xu, Z. Huang and L. Zhang, Green Chem., 2024, 26, 2922–2927.
80. Z. Wang, J. Zhai, Y. Zhao, W. Wang, Z. Lu and P. He, J. Catal., 2024, 431, 115409.
81. I. Deroche, T. J. Daou, C. Picard and B. Coasne, Nat. Commun., 2019, 10(1), 4642.
82. Z. Wang, Y. Zhao, K. Liu, J. Li, Y. Zhang, D. Wang, P. He and Z. Zhang, ACS Catal., 2025, 12157–12167.
83. N. Yang, H. Li, X. Lin, S. Georgiadou, L. Hong, Z. Wang, F. He, Z. Qi and W. F. Lin, Elsevier B.V., 2024, vol. 12, p. 073.
84. Q. Ji, B. Tang, X. Zhang, C. Wang, H. Tan, J. Zhao, R. Liu, M. Sun, H. Liu, C. Jiang, J. Zeng, X. Cai and W. Yan, Nat. Commun., 2024, 15, 8089.
85. D. Wu, F. Jiao and Q. Lu, ACS Catal., 2022, 12(20), 12993–13020.

---

Footnote

† These authors contributed equally to this work.

---